Ann. occup. Hyg., Vol. 47, No. 1, pp. 90, 2003
© 2003 British Occupational Hygiene Society
Published by Oxford University Press
Letters to the Editor |
Reply
Department of Epidemiology, INRS, BP 23, 54501 Vandoeuvre Cedex, France
Received 29 September 2002;
The comments of Burstyn and Kromhout (2003) on our paper (Wild et al., 2000) give me the opportunity to try to clarify some of the issues. Burstyn and Kromhout remark that our approach is similar to that used by Ramachandran (2001), which it is indeed, with the difference that the experts were used to group the different exposure groups into a limited number of semi-quantitative ratings, rather than to explicitly model the process by which the exposure arises. Burstyn and Kromhout find it odd that the raters should be ignorant of the exposure measurements and suggest, in substance, that an expert without knowledge of the exposure measurements has no idea of the exposure level.
Despite their impressive literature citations to prove their point, I do believe that experienced industrial hygienists are in most cases able to assess semi-quantitatively, although perhaps imprecisely, the exposure levels. This assessment is usually based on their knowledge of the industrial process and other determinants of the exposure—Cherrie and Schneider (1999), for instance, give a structured way to obtain such assessments—but also, and perhaps more importantly, on interviews with the workers (or former workers) themselves. This approach worked in the example shown in the paper and in an earlier work in which such expert ratings (obtained independently from exposure measurements) correlated with the exposure measurements (Moulin et al., 1997), although admittedly weakly, and allowed us to show a dose–response effect of exposure to hard-metal dust and lung cancer (Moulin et al., 1998; Wild et al., 2000). For workplaces for which measurements exist in reasonable numbers, our examples show that they will prevail, the problem is for the workplaces for which few or no measurements exist, in which case our procedure allows the combination of the expert assessment with whatever measurements do exist. Now, ‘Who qualifies to be an expert?’ My answer is that an expert is anybody (or any group) who is capable of synthesizing what can be known about the exposure determinants and ancillary information from the workers themselves to obtain semi-quantitative estimates. The quality of these estimates is then easily seen through our model of these data. If the experts are bad, there will be no relationship between measurements and the expert rating, and the between exposure group variance will be high. In this case the experts will de facto provide no quantitative information and you will know it. If you had provided the experts with the measurements in the first place, I doubt that the exposure assessments would have been very different for the EGs without measurements, but you would not know that the assessment is bad. A final point is that I do not advocate the use of experts (except maybe in approaches like Ramachandran’s) when exposure determinants are well known, but in the words of Burstyn and Kromhout (2002), ‘the information on determinants of exposure is usually harder to obtain than measurements’.
REFERENCES
Burstyn I, Kromhout H. (2002) Letter to the editor. A critique of Bayesian methods for retrospective exposure assessment Ann Occup Hyg; 46: 429–32.
Burstyn I, Kromhout H. (2003) Letter to the editor. Who qualifies to be an expert? Ann Occup Hyg; 47; 89.
Cherrie JW, Schneider T. (1999) Validation of a new method for structured subjective assessment of past concentrations. Ann Occup Hyg; 43: 235–45.
Moulin J-J, Romazzini S, Lasfargues G et al. (1997) Elaboration d’une matrice emplois-exposition dans l’industrie productrice de metaux durs. Rev Epidemiol Santé Pub; 45: 41–51.[Web of Science][Medline]
Moulin J-J, Wild P, Romazzini S et al. (1998) Lung cancer risk in hard metal workers. Am J Epidemiol; 148: 241–8.
Ramachandran G. (2001) Retrospective exposure assessment using Bayesian methods. Ann Occup Hyg; 45: 651–67.
Wild P, Perdrix A, Romazzini S, Moulin J-J, Pellet F. (2000) Lung cancer mortality in a hard metal. Occup Environ Med; 57: 568–73.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||